The Limitations of Promising Pitch Comps

I’m never really sure how much to review recent articles. Last week I wrote a lot about pitch comps. Compared Henderson Alvarez‘s changeup to Felix Hernandez‘s changeup. Compared guys to certain signature pitches like Aroldis Chapman‘s fastball and Clayton Kershaw‘s slider. Wound up with this Marcus Stroman absurdity, in which Stroman is flattered by some of the strongest comps in the game. The basic idea: Stroman’s regular fastball shares the same traits with Johnny Cueto‘s. His sinker shares the same traits with Roy Halladay‘s. His curveball shares the same traits with Jose Fernandez’s. And so on. The whole idea was comparing pitches to other pitches based on average velocity, horizontal movement, and vertical movement. It’s a little informative, and sometimes a lot of fun.

I wanted to put together this follow-up, which serves to caution you not to make too much of a good comp. There’s more that goes into a pitch than how it moves, and how fast it moves. There’s definitely more than that, when it comes to how a pitch works within a given repertoire of many pitches. And, our examples: Cole Hamels and J.A. Happ. Last week, when I ran some pitch comps, I found that Happ’s changeup compares the strongest with Hamels’ changeup. For one thing, we know Happ’s changeup isn’t as good as Hamels’ changeup. For another thing, it goes deeper than this.

To nip this in the bud, how do we know Happ’s changeup isn’t as good as Hamels’ changeup? Happ is a worse pitcher than Hamels is, and Happ also has thrown his changeup less often than Hamels has. That’s basically all the proof you need. If Happ had a better changeup, he’d throw more changeups, and he’d presumably be better overall. Now let’s take this beyond just changeups.

Here are Cole Hamels’ pitches, from Brooks Baseball:

• Four-seam fastball
• Two-seam fastball
• Changeup
• Curveball
• Cutter

It’s a five-pitch repertoire, and Hamels makes plenty of use of all his weapons. He’s primarily known for his change, but one of the reasons the changeup is so effective is because hitters are left thinking about an assortment of options. Using the Baseball Prospectus PITCHf/x leaderboards, I ran some pitch-comp analysis for Hamels’ pitches, checking other left-handed starters from 2014 who threw at least 50 pitches of the given type.

Looking at Hamels’ four-seam fastball, the closest comp? J.A. Happ’s four-seam fastball, with a comp rating of 0.4. Happ, interestingly, has been gaining velocity year after year. It’s one of the reasons the Mariners sought him out. Happ’s fastball last season became Cole Hamels’ fastball.

Looking at Hamels’ two-seam fastball, or sinker, the closest comp? Happ’s two-seam fastball, with a comp rating of 1.0. It’s actually dead even with Wei-Yin Chen’s two-seam fastball.

Looking at Hamels’ changeup, well, we’ve already been over this, but, the closest comp? Happ’s changeup, with a comp rating of 0.6.

At last, looking at Hamels’ curveball, the closest comp is Jeff Locke, but Happ’s curve comes in fifth. Additionally, Hamels seems to have made some tweaks to his curve, because the movement changed a few inches from the season before. Happ’s 2014 curve was a stronger comp for Hamels’ 2013 curve, and suffice to say Hamels in 2013 was also a good pitcher.

Now, the cutters aren’t great matches. Like Hamels, Happ does throw a cutter, but it’s slower, and it has more sink. So the two pitchers aren’t completely identical. But the cutter is only part of the equation, and on balance, J.A. Happ’s repertoire is a fantastic match for Cole Hamels’. They throw the same five pitches, and four of them are strikingly similar, in how fast they fly, and in how they break. And maybe this isn’t a shock — both Hamels and Happ were developed by the Phillies, and when Happ came around, Hamels was an established front-of-the-rotation starter. I’m sure advice was requested. I’m sure advice was given. It’s not a total coincidence similar paths have led to similar styles.

Yet, Cole Hamels is regarded as an ace, while Happ simply fetched a year or two of Michael Saunders. Though Happ is useful, why does he fall short of the Hamels standard? It’s the rest of the stuff, which you can distill into the word “consistency” if you like. It’s one thing to throw similar pitches; it’s quite another to throw them similarly. Hamels and Happ have differences in their throwing motions, and they also have differences in how well they can locate. This isn’t a total surprise, but some images can help. Service was provided by Baseball Savant. For the sake of simplicity, the following images show just pitch locations against right-handed hitters.

I don’t think you see massive differences in where Hamels and Happ put their fastballs and cutters.

Happ’s gotten pretty good with his fastball, as it has gained strength. Hamels has gone in off the plate a little more often, which you’d expect from someone with a superior cutter, but for the most part, here you observe similar patterns. On the other hand, here are where curveballs went:

Remember, this is just against righties, and you can see Hamels effectively back-footing a lot of his breaking balls. Happ left more curves over the middle, and his curveball wound up away more often. By the numbers, I counted 57% of Hamels’ curveballs to righties ending up both low and inside. For Happ, I got 36%. Some of this might’ve been intent; a lot of this, probably, was worse command.

And, the changeups:

That ought to be pretty telling. Against opposite-handed hitters, a pitcher almost always wants to put his changeup somewhere in the low-away quadrant. Both pitchers succeeded in throwing a little over 70% of their changeups to righties away. But, while Happ threw 41% of his changeups to righties away and below the midpoint of the strike zone, Hamels came in at 59%. Happ had more trouble not flying open, and he struggled, relatively speaking, to keep the changeup by the knees. This explains why Hamels threw one changeup for every four pitches to a righty, while Happ was at one out of nine. Hamels threw the pitch with a lot more confidence. With two strikes, the difference was more stark: Hamels threw 29% changeups, and Happ threw 8%.

To sum up what might not need summing up: J.A. Happ is a worse starting pitcher than Cole Hamels is. Yet, Happ throws very similar pitches to Hamels’ pitches. The cutters are different, but everything else is remarkably close, and so where do the differences come from? Hamels is just more consistent in achieving the right throwing motion, while Happ wanders a bit too much, in particular with his curve and his change. When Happ is really on his game, he’ll look a lot like a dominant Cole Hamels. Perhaps a dominant Cole Hamels who throws a few more fastballs than usual. But consistency is almost always the difference between being sometimes good and being frequently good, and though the comps make Happ into an encouraging sleeper, he’s also 32. So.